Abstrakt
The aim of this paper is to introduce a spoken corpus of Czech intermediate learners of English. The poster will present the structure of the corpus, will provide the summary of the metadata of 50 students recorded for this corpus and will also introduce the challenges faced while designing such a corpus.
The corpus is being built and since a corpus of advanced learners of English (LINDSEI) already exist, the aim of this corpus was to provide comparable data. When planning the design of the corpus, the previous corpus was taken into account.
The three tasks (speaking about one of three topics, dialogue and picture description) remained but were slightly altered to be suitable for intermediate learners of English - a new picture was chosen and the topics were adjusted. Based on our experience with the existing corpus, two tasks were added - students were asked to choose one topic and talk about it in Czech and the recording session started with a reading task in English.
There were several challenges met while building the corpus: how to establish the proficiency level in advance? How to make the students that do not know the interviewer more comfortable? How long should an interview be? The poster will try to answer these questions and present the problems with establishing the proficiency level instituonally, with putting the tasks into a specific order and with trying not to make the interview too long.